Image speed estimation is used in a variety of visually mediated behaviors, from path integration to navigation. From studies in both vertebrates and invertebrates it appears that the behavioral, organizational and computational principles of visual motion detection are highly convergent across taxa. While many of these systems are well understood, the neural basis of speed estimation remains a mystery. This proposal will investigate the possibility that non-directional motion units are used in the brains of honeybees to estimate image speed. The aims of this study are to: test and develop more advanced non-directional models of speed estimation, compare them to alternative models of speed and velocity estimation in primates, mimic behaviors dependent on visual speed estimation in simulation, and test the model in a real world environment using an autonomous robotic system. This will be accomplished by drawing from the large body of anatomical, electrophysiological and behavioral data on motion detection in insects to create computer simulations and apply the algorithms to artificial systems. These simulations will then be used to make testable hypotheses and direct our collaboration with electrophysiologists and anatomists. [unreadable] [unreadable] [unreadable]